A Comparison of Salary Determination in the Free Agent and Salary Arbitration Markets
Daniel R. Marburger
Salary arbitration has existed as a means of resolving salary disputes in major league baseball since 1974. In the event a player and his club reach an impasse in salary negotiations, each side submits an offer to an independent arbitrator, who must then select one of the offers as the binding settlement.
As with similar arbitration procedures (generally referred to as "final-offer arbitration"), the purpose of baseball's salary arbitration is not to have an arbitrator determine the players' salaries, but rather for arbitration to serve as a threat to continued impasse. The assumption is that the parties would prefer negotiating their own settlement to having one imposed by a third party. In this regard, salary arbitration in baseball seems to have been effective: since its inception, nearly 80% of those filing offers negotiated a settlement without resorting to arbitration ( Fizel 1994).
Despite the propensity for baseball's salary arbitration to generate negotiated settlements, Crawford ( 1979) warned that arbitration avoidance did not necessarily connote an effective procedure. Should the arbitrator essentially dictate the settlement, leaving the bargainers with little room to negotiate on their own, then arbitration will not fulfill its objective.
Theoretical models of final-offer arbitration suggest that the preferences of the arbitrators drive not only the submitted offers but negotiated settlements as well ( Farber 1980, Farber and Katz 1979). Due to the critical role that arbitrator preferences play in determining settlements, the source of those preferences needs to be identified to fully evaluate the procedure.
In baseball, comparative baseball salaries are listed in the collective bargaining agreement as permissible evidence in arbitration hearings. Given that baseball's free agent market begins two months prior to the arbitration filing period, one might expect that free agent salaries would serve as the basis for arbitrator preferences. Curiously, however, Marburger ( 1994) found that the